Constant pH tutorial

User discussions
1

GROMACS version: 2021-beta1-dev-UNCHECKED
GROMACS modification: Yes

Hi,

I’m trying to gain facility with the new update of GROMACS that allows for constant pH simulations. I am following the tutorial on the Github page (1_ASP_simulations/tutorial.md · main · Gromacs constant ph / tutorials · GitLab).

At the step where you parameterize the V_MM potential, I am getting a good agreement for the Asp potential 7th order fit but NOT for the buffer. The tutorial states “Coefficients for buffers can be obtained with the exact same procedure we applied for ASP.” However, this is unclear to me and I’m hoping someone can point me to what I’m doing wrong.

I can calculate the average value of the dv/dlambda for the Asp residues by averaging the second column of the cphmd-dvdl-1-2.xvg file. I am trying to do a similar procedure to obtain the averages for the third column to get the buffer potential. However, my averages are quite far from the ones the tutorial obtains. For example, my intercept of the 7th-order fit is 715 instead of 575 (in the tutorial) and this makes a big difference when actually simulating the model.

I also did not understand where the Note about creating an ion of type BUF1 in ions.itp is relevant. Is there supposed to be a different simulation to be run for parameterizing the Buffers?

Any help is appreciated!

2

Hi!

You should indeed parameterise the buffer particle on its own (or use the coefficients we provide), not with the same parameterization simulation as the Asp. For the buffer parameterization we provide the input files (a system of two buffer particles) in the Supplementary Information of the original paper regarding constant pH MD (https://doi.org/10.1021/acs.jctc.2c00516). In that system, the buffer has the correct charge range -0.5 … +0.5.

In the system for Asp parameterization, one buffer particle is added to the system simply because we want to account for the charge upon deprotonating Asp. This buffer particle (could be any ion as well) has the opposite charge as Asp (hence when in parameterization for example Asp has -1 charge at lambda=1, the buffer in that system has charge +1). This is is because one should not run with a non-neutral simulation box. During parameterization simulations, we also apply position restraints in order to keep the Asp and Buf far from each others.

Best,
Noora Aho

1 Like